Collective Choice Lotteries: Dealing with Randomization in Voting, Matching, and Allocation

点击翻译按钮获取中文摘要

Randomization is playing an ever increasing role in collective choice with examples ranging from voting to matching markets and fair allocation. As it turned out, randomization not only enables ex ante fairness, but also the circumvention of classic impossibility results such as the theorems by Arrow and Gibbard-Satterthwaite. These observations are important for human societies as well as computational multiagent systems which draw on collective choice rules for decision-making. The goal of this project is to provide a comprehensive treatment of randomization by studying the following interdisciplinary and interrelated research questions: (i) when are collective choice lotteries acceptable, (ii) how do agents compare lotteries, and (iii) how can randomized rules be implemented. It is planned to address these questions using formal approaches from mathematical economics and theoretical computer science as well as experiments involving human subjects. Particular emphasis will be put on the individual peculiarities of the application domains under consideration.